Fix wrong comparison of embeddings

[nachoalonsoportillo]

Issue Overview

The embeddings being compared are those of answer given in code as "3rd century" with those of answer provided by model for question "When was tea discovered?", which is correct, and those of answer provided by model for question "When was tea discovered?" with itself, which is incorrect.

Describe your environment

Steps to Reproduce

1. Run code in branch 05_03.

Expected Behavior

Current Behavior

Possible Solution

Replace this code

building-apps-with-ai-tools-chatgpt-semantic-kernel-langchain-4469616/src/main.py

Line 35 in fc71a7c

for i in range(0, len(question_answers)*2, 2):

with this code:

for i in range(0, len(question_answers)):

And this code

building-apps-with-ai-tools-chatgpt-semantic-kernel-langchain-4469616/src/main.py

Line 37 in fc71a7c

embedding_b = resp['data'][len(question_answers)]['embedding']

with this code:

embedding_b = resp['data'][i+len(question_answers)]['embedding']

Screenshots / Video

Related Issues